coloring 3D maps according to local cross-correlation fit of docked PDB structures
Dear users, we would like to color a Cryo-EM map according to the cross-correlation fit of docked PDB structures. Is that possible in Chimera ? Best wishes J-F Menetret
Hi J-F, Do you mean if the correlation of a molecular model with a map is 0.9 then color the whole map blue? Or do you want to color just the region of the map near the molecule blue. Or do you wan the map be colored in different colors, each representing the local correlation of the molecule to the map near the point where that color is shown? If this last case, how would you define the "local correlation"? Probably you would want to color the molecule in that case not the map since it would give a better view of the whole volume rather than just its envelope. Tom -------- Original Message -------- Subject: [Chimera-users] coloring 3D maps according to local cross-correlation fit of docked PDB structures From: Jean-Francois Menetret To: chimera-users@cgl.ucsf.edu Date: 10/18/12 6:24 AM
Dear users, we would like to color aCryo-EM map according to the cross-correlation fit of docked PDB structures. Is that possible in Chimera ? Best wishes J-F Menetret
Hi J-F, I added the "vop localCorrelation" command to Chimera that computes the correlation between two maps over a moving window of size N by N by N voxels. The result can be used to color surfaces or atoms as shown in the attached picture of a zinc transporter (3j1z) and 13 Angstrom resolution EM map (EMDB 5450). To color the surface I used open 3j1z molmap #0 13 open emdbID:5450 vop localCorrelation #0 #1 windowSize 5 model #2 scolor #1 volume #2 cmap .8,blue:0.9,white:1,red transparency 30 #1 If you leave off the windowSize option to vop localCorr it defaults to 5 voxels. The colored atoms in the picture and the color key were made with the Values at Atom Positions dialog (menu Tools / Volume Data / Values at Atom Positions) and the Render by Attribute dialog (menu Tools / Depiction / Render by Attribute). The calculation is quite slow because it is being done in Python. It took a minute for the 28 x 32 x 37 simulated map and 160 x 200 x 160 EM map. It computes local correlation at the grid points of the first map in the vop localCorr command. It will only be a little effort to do the calculation in C++ which will be one hundred times faster but I don't think I have time to do that today. The new command will be in tonight's daily build. Tom -------- Original Message -------- Subject: Re: [Chimera-users] coloring 3D maps according to local cross-correlation fit of docked PDB structures From: Jean-Francois Menetret To: Tom Goddard Date: 10/19/12 8:35 AM
Hi Tom,
we would like tocolor the surface representation of our molecule with the local correlation score between the fitted atomic models and the EM-maps; you can find an example in figureS3 of Hipp et al. Nucleic Acids Research, 2012, Vol. 40, No. 7 3275--3288
best wishes J-F
On Thu, Oct 18, 2012 at 7:37 PM, Tom Goddard wrote:
Hi J-F,
Do you mean if the correlation of a molecular model with a map is 0.9 then color the whole map blue? Or do you want to color just the region of the map near the molecule blue. Or do you wan the map be colored in different colors, each representing the local correlation of the molecule to the map near the point where that color is shown? If this last case, how would you define the "local correlation"? Probably you would want to color the molecule in that case not the map since it would give a better view of the whole volume rather than just its envelope.
Tom
-------- Original Message -------- Subject: [Chimera-users] coloring 3D maps according to local cross-correlation fit of docked PDB structures From: Jean-Francois Menetret To: chimera-users@cgl.ucsf.edu <mailto:chimera-users@cgl.ucsf.edu> Date: 10/18/12 6:24 AM
Dear users, we would like to color aCryo-EM map according to the cross-correlation fit of docked PDB structures. Is that possible in Chimera ? Best wishes J-F Menetret
Hi J-F, I replace the Python local correlation calculation with C++ so it is relatively fast now -- tens seconds using two 256 by 256 by 256 maps with window size 5 pixels. Tom
Hi J-F,
I added the "vop localCorrelation" command to Chimera that computes the correlation between two maps over a moving window of size N by N by N voxels. The result can be used to color surfaces or atoms as shown in the attached picture of a zinc transporter (3j1z) and 13 Angstrom resolution EM map (EMDB 5450). To color the surface I used
open 3j1z molmap #0 13 open emdbID:5450 vop localCorrelation #0 #1 windowSize 5 model #2 scolor #1 volume #2 cmap .8,blue:0.9,white:1,red transparency 30 #1
If you leave off the windowSize option to vop localCorr it defaults to 5 voxels. The colored atoms in the picture and the color key were made with the Values at Atom Positions dialog (menu Tools / Volume Data / Values at Atom Positions) and the Render by Attribute dialog (menu Tools / Depiction / Render by Attribute).
The calculation is quite slow because it is being done in Python. It took a minute for the 28 x 32 x 37 simulated map and 160 x 200 x 160 EM map. It computes local correlation at the grid points of the first map in the vop localCorr command. It will only be a little effort to do the calculation in C++ which will be one hundred times faster but I don't think I have time to do that today.
The new command will be in tonight's daily build.
Tom
-------- Original Message -------- Subject: Re: [Chimera-users] coloring 3D maps according to local cross-correlation fit of docked PDB structures From: Jean-Francois Menetret To: Tom Goddard Date: 10/19/12 8:35 AM
Hi Tom,
we would like tocolor the surface representation of our molecule with the local correlation score between the fitted atomic models and the EM-maps; you can find an example in figureS3 of Hipp et al. Nucleic Acids Research, 2012, Vol. 40, No. 7 3275--3288
best wishes J-F
On Thu, Oct 18, 2012 at 7:37 PM, Tom Goddard wrote:
Hi J-F,
Do you mean if the correlation of a molecular model with a map is 0.9 then color the whole map blue? Or do you want to color just the region of the map near the molecule blue. Or do you wan the map be colored in different colors, each representing the local correlation of the molecule to the map near the point where that color is shown? If this last case, how would you define the "local correlation"? Probably you would want to color the molecule in that case not the map since it would give a better view of the whole volume rather than just its envelope.
Tom
-------- Original Message -------- Subject: [Chimera-users] coloring 3D maps according to local cross-correlation fit of docked PDB structures From: Jean-Francois Menetret To: chimera-users@cgl.ucsf.edu <mailto:chimera-users@cgl.ucsf.edu> Date: 10/18/12 6:24 AM
Dear users, we would like to color aCryo-EM map according to the cross-correlation fit of docked PDB structures. Is that possible in Chimera ? Best wishes J-F Menetret
_______________________________________________ Chimera-users mailing list Chimera-users@cgl.ucsf.edu http://plato.cgl.ucsf.edu/mailman/listinfo/chimera-users
participants (2)
-
Jean-Francois Menetret -
Tom Goddard